Intel made the 2022 release of its oneAPI toolkit available for processor extensions to DevOps teams.
The goal is to make it simpler for DevOps teams to expose extensions to any company’s processors at runtimes that developers use, said James Reynders, Intel’s chief evangelist for software products.
At its core, the oneAPI Toolkit is a set of Intel-built APIs for x86, graphics processing units (GPUs), application-specific integrated chips (ASICs), field-programmable gate arrays (FPGAs), and other classes of processors.
Reinders pointed out that developers have never had ready-made access to many different classes of processors before. However, the challenge for organizations is to make it easier for developers to take advantage of these capabilities in a way that doesn’t require them to write low-level code.
The OneAPI 2202 toolkit adds additional capabilities including the world’s first unified compiler using C++, SYCL and Fortran, data paralleling Python for CPUs and GPUs, advanced accelerator performance modeling and performance tuning and acceleration for artificial intelligence (AI) and visualization workloads ray tracing. Collectively, Intel said it made 900 new and improved features available via the oneAPI toolkit in the past year, including deeper Microsoft Visual Studio Code integration, support for Microsoft Visual Studio 2022 and Microsoft WSL2 for Linux development on Windows.
As information technology continues to evolve, it will become more common for developers to invoke heterogeneous classes of processors within the same application, Reynders says. These capabilities, he added, will lead to an explosion of software development innovations during the rest of the decade and beyond.
Reynders said Intel is specifically focused on making it easier for developers to call inference engines that take advantage of processor extensions to run AI workloads. Going forward, most applications will make use of machine learning algorithms that run Intel-optimized inference engines via extensions to their processors. Reynders noted that as these workloads are created, integrating AI models into DevOps workflows will become the next major challenge.
It is too early to say how simplified access to heterogeneous processors, also known as XPUs, might change the application development landscape. One thing that is clear is that IT organizations now have more options than ever before. The problem now is to find a way to invoke all of these capabilities in a way that doesn’t confuse the developers. DevOps teams will play a key role in deciding which processors and associated add-ons to make available to their development teams.
In the meantime, the battle for control of the processor will intensify in the coming months. DevOps teams will need to determine whether they are comfortable using processors made by one company such as Intel or whether they want to use a mix of processors provided by multiple vendors. Whatever they decide, it looks like APIs will play an important role in making it easier to take advantage of advances in processor technologies.